Gender equality in the name of the state: state feminism or femonationalism in civic orientation for newly arrived migrants in Sweden?

This article contributes to ongoing discussions in the social sciences about how to interpret the incorporation of gender equality into integration policies – is it a form of state feminism or femonationalism? Drawing upon intersectionality, we analyse how gender equality is presented, discussed and negotiated in relation to ethnicity and nationality in Sweden. Methodologically, we employ a bifocal lens that combines (1) a quantitative investigation of representations of civic orientation programmes in Swedish policy documents and mainstream media, and (2) a qualitative analysis of ethnographic data collected in six civic orientation courses – three in English and three in Arabic – in three large municipalities. Such a two-pronged approach, which connects policy and media discourses with interactions in civic orientation classes, o ﬀ ers a granular picture of the complex and often ambivalent intersections of ethnicity and gender in relation to migration in Sweden. Ultimately, the co-optation of feminist values brings with it the risk of warping feminism into a trait of national/ethnic distinctiveness. Crucially, femonationalism is not the prerogative of far-right parties but is already becoming institutionalised, informing both mainstream media and educational practices in a feminist state like Sweden.


Introduction
In an op-ed article in one of Sweden's most widely read dailies, then-Social Democratic Prime Minister Stefan Löfvén commented on ethnic segregation in Sweden, stating: 'It must be clear which rules apply in Sweden and which values should permeate our society … Sweden is a modern and free country where justice, fairness and gender equality are central values.They are the result of decades of struggle, and we must take them forward, we will never back down from them' (Aftonbladet, 2016).Pronouncements about the relationship between gender equality and state governance are not new or unique to Sweden.In fact, gender equality has become increasingly invoked in debates about policy proposals that supposedly enhance the 'integration' of migrants in different European contexts.Interestingly, social scientists have provided very different interpretations of the invoking of gender equality in integration debates and policies.
On one hand, Joppke (2007) underscores how the shared focus on rightsparticularly women's rightsin integration policies in Europe is indicative of a post-national turn characterised by a decrease in national and cultural distinctiveness and a convergence around universal liberal values.In Joppke's view, these policies 'are not born of sources extrinsic to liberalism, such as nationalism and racism, but are inherent in liberalism itself ' (2007, p. 14).From this point of view, it could be argued that the importance of women's rights in civic courses for migrants is a facet of what Hernes (1987) famously called 'state feminism', a well-meaning pursuit on the part of a state to improve women's lives in a polity by making gender equality the central principle that informs all policy making and practices.
On the other hand, Farris (2017) contends that 'the concrete national articulation of the themes of gender equality and women's rights within the civic orientation national programmes is precisely what attests to the persistence and even strengthening, rather than a disappearance, of a nationalist (and racist) trope, which I conceive as intrinsic and not extrinsic to liberalism ' (2017, p. 81).Farris goes on to propose that 'civic integration policies are arguably the most concrete and insidious form of femonationalism' (ibid.).Here, femonationalism refers 'both to the exploitation of feminist themes by nationalists and neoliberals in anti-Islam (but, as I will show, also anti-immigration) campaigns and to the participation of certain feminists and femocrats in the stigmatisation of Muslim men under the banner of gender equality' (Farris, 2017, p. 4).
Taking an intersectional perspective on the mutual constitution of gender, ethnicity and nationalism (see e.g.Yuval Davis, 1997), the present article seeks to contribute to these ongoing discussions by analysing how gender equality is presented, discussed and negotiated in relation to migrants' 'integration' into Sweden.For this purpose, we focus on courses in civic orientation for newly arrived adult migrants.According to the main policy document governing these courses, the aim of civic orientation is 'for participants to develop knowledge about human rights and fundamental democratic values, the rights and obligations of the individual in general, how society is organised, and practical everyday life' (SFS, 2010).Moreover, the policy document also underscores that civic orientation should be characterised by 'dialogue and respect' (SFS, 2010).Municipalities are responsible for the organisation of these courses, which are given in a language that a newcomer knows well.In terms of content, civic orientation covers eight main content areas: (1) arriving in Sweden, (2) living in Sweden, (3) supporting oneself and developing in Sweden, (4) the individual's rights and obligations, (5) starting a family and living with children in Sweden, (6) having an influence in Sweden, (7) looking after your health in Sweden,and (8) ageing in Sweden (SFS, 2010).The courses are led by 'society communicators' (samhällskommunikatörer) who usually have different educational backgrounds or qualifications, and are employed mainly because of their own experience of migration, their multilingual repertoire, and their knowledge of Swedish society.Unlike Farris's (2017) and others' investigations of civic orientation in other European countries (see e.g.Blankvoort et al., 2023), which are mainly based on qualitative analyses of policy documents, teaching materials and interviews with key stakeholders, the present article employs a bifocal lens that combines (1) a quantitative investigation of representations of civic orientation programmes in Swedish policy documents and mainstream media, and (2) a qualitative analysis of ethnographic data collected in six civic orientation coursesthree in English and three in Arabicin three large municipalities in Sweden over a period of four months in 2020.Such a two-pronged approach, that connects policy and media discourses with interactions in civic orientation classes, offers a granular picture of the complex and often ambivalent intersections of ethnicity and gender in relation to migration in Sweden.
We start with a brief summary of research about gender and migration in Sweden; we then move on to present the analytical and methodological framework that informs the article, before delving into an analysis of policy documents, media discourses and classroom interactions.We conclude with a short reflection about the role of critique from within feminism.

Gender equality in Sweden
Sweden is often described as having one of the highest degrees of equality between men and women of any country in the world (see Borchorst & Siim, 2008).Moreover, gender equality is a central component in Sweden's nation-branding (Jezierska & Towns, 2018).However, as Martinsson et al. (2016) warned, 'the equality mantra … recreates a hierarchical order between an imagined modern, highly developed "we" and a less developed 'other' that lacks those attributes ' (2016, p. 6).In other words, the discursive construction of Sweden as one of the most gender equal countries works on different scales: on a global scale, it positions Sweden as a positive 'exception', one that should be imitated by other nation-states; on a national scale, it produces a dichotomy between Swedes and migrants.A common feature of these two scales is the superiority of the Swedish 'we' on an imaginary gender equality ranking.
How such a 'myth of Swedish gender equality' (Martinsson et al., 2016) materialises discursively has been put under close scrutiny in a large body of scholarship informed by intersectionality (see, e.g.Martinsson et al., 2016;Andersson, 2018;Lane & Jordansson, 2020).Such a theoretical framework has made it possible to bring into relief the ethnic and racial colorations of Swedish state feminism, thus unveiling the underlying 'politics of gendered and racialized divides' (de los Reyes and Mulinari, 2020) upon which integration policies are built.
For example, in a recent overview of Swedish policy documents about social welfare, family life and the labour market over a period of 40 years, Mulinari (2021) illustrated how the discursive category 'migrant woman' (invandrarkvinna) has shifted meaning over the years.In the 1970s and 1980s, policy documents were characterised by the co-presence of (1) a dialogue with migrant organisations and 'inclusion, respect and protection of what was defined as "the migrant family"' (2021, p. 189), and (2) the representation of the migrant woman as located in the 'home', isolated from society, and therefore in need of being liberated by the state.The latter discursive construction became entrenched in the following decades in connection with the emergence of a discourse of cultural distance, one that highlights the difference in the value system between migrants and Swedes.While the word values (värderingar) is used in these policy documents in relation to Swedes, cultures (kulturer) is employed for migrants.Such a value/culture chasm is employed to legitimate the need for intervention in order to 'save women and girls and the "second generation" ' (2021, p. 193); that is, migrants' offspring (of both genders).
The main textbook for civic orientation for adult migrants includes a similar opposition between Sweden as a feminist state informed by the principle of gender equality and a rather undifferentiated group of migrants who need to be educated (if not rescued and saved).As Carlson et al. (2021) illustrated through close textual analysis, the textbook represents the historical process that led to universal voting rights and to gender mainstreaming in Sweden as 'effortless ' (p. 203).Moreover, the lack of comparison with historical developments in other national contexts makes Sweden appear 'unique'.Even in those few instances where other contexts are mentioned, such as in the case of honour-related violence, these societies are invoked in order to highlight 'migrants' presupposed patriarchal customs ' (2021, p. 204) and thereby justify the need for migrants 'to alter in order to fit into Swedish society ' (2021, p. 205).
Crucially, migrant women are not the only object of concern in Swedish political debates; Swedish women have also been the target of political interventions.In policy documents, Swedish women are represented in relation to gender mainstreaming and increased participation in the labour market (Mulinari, 2021).Moreover, they have more recently become the 'object of care'obsession evenof the supporters of Sweden Democrats, a far-right party with Nazi roots that has enjoyed increasing electoral success over the years.While the party has overtly anti-feminist and anti-immigrant agendas (Mulinari & Nergaard, 2017), some feminist ideas about gender equality have been 'co-opted to argue against migration and especially in order to demonise migrant men' in order to justify the party's call for measures that would enhance the safety and security (trygghet) of Swedish women in particular, and Swedish society at large.
While the notion of femonationalism has been employed to capture the strategic cooptation of some feminist values by far-right parties in Sweden (Sager & Mulinari, 2018), it has not been used to theorise the benevolent mainstream discursive construction of dichotomies between more patriarchal migrant men and oppressed migrant women, on one hand, and more gender-equal Swedish men and women, on the other, both in policy documents and teaching materials for adult migrants.While this could be due to the fact that femonationalism is a fairly recent concept, we want to caution against the possible effects of isolating femonationalism as the ideological underpinning of farright parties alone.Such a discursive move may inadvertently lead to (1) positioning far-right political formations and their supporters as aberrant fringes outside the political mainstream, which could lead to (2) locating femonationalism outside the workings of existing state apparatuses.

Theory and method: intersectionality and discourse ethnography
We take an intersectional perspective to critical discourse analysis (see also Kitis et al., 2018) that brings together the quantitative approach of corpus-assisted discourse studies with qualitative insights offered by ethnographic fieldwork.Intersectionality has established itself as a paradigm through which to 'reveal how power works in diffuse and differentiated ways through the creation and deployment of overlapping identity categories' (Cho, Crenshaw & McCall, 2013, p. 797).Through imbrication, categories mutually constitute one another in such a way that, say, 'class, race, and gender do not exist as entities unto themselves.Instead, they crucially depend for their meaning on their relationship to the other categories with which they intersect' (Levon, 2015, p. 298).Analytically, intersectionality can be operationalised by 'asking the other question,' as Matsuda (1997) put it.Such an exercise can be summarised as follows: 'When I see something that looks racist, I ask 'Where is the patriarchy in this?' When I see something that looks sexist, I ask 'Where is the heterosexism in this?' (Matsuda, 1997, p. 66).Or, in the case of the present article, 'When I see gender equality, where is nationalism in this?' While intersectionality has typically been associated with qualitative techniques and has been put to work in analyses of small samples of texts, Baker and Levon (2016) have argued that there is also some value in investigating the mutual constitution of identity categories with the help of quantitative corpus-assisted methodologies (see also Kitis et al., 2018), which makes it possible to chart broader discursive trends on the basis of larger sets of textual data.In the present article, the bringing together of quantitative analysis of media texts and qualitative ethnographic insights of civic orientation classes is inspired by the discourse ethnographic approach proposed by Krzyzanowski, 'which integrates anthropological and critical-analytic perspectives through, on the one hand, extensive fieldwork and ethnography in institutional spaces and, on the other hand, the critical analysis of discourses of (social) actors' shaping those spaces and acting therein ' (2011, p. 235).Through a discourse ethnographic approach, it is possible to first map whether policy documents and media texts discursively construct intersectional identity bundles (for instance, we/gender equal/ethnically Swedes versus them/gender unequal/migrants), and then investigate whether, and if so, how, these intersectional nexus points are reproduced and/or contested in civic orientation classes.
The corpus of policy documents consists of two texts that have been key to the establishment and development of civic orientation: (1) the directive with which the government appointed an investigator to develop a comprehensive proposal for how civic orientation should be organised, and which topics it should cover (Dir., 2009); and (2) the investigator's report (SOU, 2010).To build the media corpus, we used the online news database Mediearkivet and searched for articles in mainstream national dailies containing the word samhällsorientering (civic orientation) from 1 January 2009 to 30 October 2021.The resulting corpus contains over 400 articles, totalling 304,505 words.The freeware corpus tool LancsBox was used to run collocate and concordance searches, focusing mainly on the terms 'gender equality' (jämställdhet), 'woman/women' (kvinna/kvinnor) and 'value(s)' (värde(n)/värderingar).However, in order to fully investigate the gender dynamics presented in the policy documents, we also considered the words 'man/men' (man/män) and 'husband/husbands' (make/makar).
Collocation refers to words that co-occur in a non-random way in language use and can help us unveil semantic prosodies; that is, the 'consistent aura of meaning' (Louw, 1993, p. 157) that a word accrues over time.For example, 'if illegal and immigrant are often paired as collocates, we may be primed to think of one concept even when the one is not present' (Gabrielatos & Baker, 2008, p. 11).As such, collocates can be useful to tease out which topoithat is, 'the common-sense reasoning typical for specific issues' (Van Dijk, 2000, p. 97-98) are used in order to justify particular political interventions.
This is because semantic prosodies encoded in collocates 'can help create, reinforce, or exploit a topos without the need for any explicit argumentation' (Gabrielatos & Baker, 2008, p. 22).We ran collocate searches for each search term separately (both singular and plural forms), with a span of five words to the left and five words to the right of the search term.We ordered collocates according to their MI3 (mutual information) statistical score (cut-off point: MI3 ≥ 12 and a frequency ≥ 5).Further analysis was conducted by examining concordanceslines in tabular form that show occurrences of a word or phrase in its immediate co-textor entire newspaper articles.
The ethnographic data analysed in this article consists of fieldnotes taken during participant observation of six civic orientation coursesthree in English and three in Arabicin the spring of 2020.We followed the principles of research ethics spelled out by the Swedish National Research Foundation (Vetenskapsrådet).All participants gave written consent to participate in the study and all the names used in this article are pseudonyms.The fieldwork led to 600 pages of notes, which were then coded according to themes such as 'age/ageing', 'food', 'gender', 'health', etc.The main reason for choosing these courses is that the majority of migrants in Sweden are Arabic speakers and Arabic is the largest minority language in Sweden.Moreover, those who attend the English-speaking courses originate from countries including Bangladesh, Cameroon, Pakistan, Peru, and Vietnam.By focusing on civic orientation classes in Arabic and English, it was possible to gain insight into multiple migrant perspectives.As such, our analysis is not generalisable to civic orientation as a whole.Rather, it is a 'case study'; it provides a window into how civic integration policies are put to work in practice in Sweden.

Policy documents and media representations about civic orientation: gender equality, women and values
The government's directive, establishing an official inquiry into civic orientation, only makes one mention of gender equality, albeit a significative one, in a section listing the topics to be covered in civic orientation: Excerpt 1 Civic orientation should (ska) convey the importance of fundamental values such as democracy and the equal value of all human beings.It ought to (bör) include knowledge acquisition and reflection about what it means to live in Swedish society, gender equality and respect for young people's integrity.It should (ska) also provide clear information about the rights and obligations of the individual.Civic orientation ought to (bör) provide space for dialogue and reflection.(Dir., 2009) It is not particularly controversial for a democratic state like Sweden to emphasise fairness and equal rights and duties as key issues for migrants to be acquainted with.Moreover, this excerpt emphasises that civic orientation ought to be a platform where migrants engage in 'dialogue and reflection' about Swedish society.However, the repetition of markers of strong deontic modality (ska and bör) indicates a normative stance on the part of the government about which topics are not negotiable and therefore should be included into this educational provision, such as gender equality and respect for 'young people's integrity' (ungdomars integritet).As Fairclough (2003, p. 173) has pointed out, statements with deontic modalities are imbued with more or less implicit, assumed evaluations.In the case of the extract above, the assumption is that migrants lack knowledge of gender equality and young people's rights, so it is imperative for them to learn and reflect about them in civic orientation classes.
A similar representational pattern in which gender equality is tightly tied to other general civic values (such as fairness and youth's rights) also emerges if we look at the collocates of 'gender equality' ( jämställdhet) in the media texts.As mentioned earlier, collocates are words that are employed together with a frequency greater than chance.Table 1 below presents the strongest collocates of gender equality, grouped into semantic categories.
To begin with, it is unsurprising that the Liberal Party and one of its leading figures, Nyamko Sabuni, feature as strong collocates of gender equality.This can be partly explained as a result of the fact that Sabuni was minister of integration and gender equality when civic orientation was established.Even more interesting is how, analogous to Excerpt 1 above, gender equality strongly co-occurs with a plethora of civic principles and values such as democracy, freedom, human rights, the rights of the individual, and youth's integrity.While space limitation prevents us from reproducing here all the concordances of the word jämställdhet in the corpus, a closer look at its context of usage shows how gender equality seldom appears on its own but is part of word clusters containing one or several of these civic principles (such as 'gender equality and democracy'), which are considered necessary to teach to migrants.Overall, the collocates of 'gender equality seem to indicate the presence of a topos of usefulness/advantage according to which: if an action under a specific relevant point of view will be useful, then one should perform it […] To this topos belong different subtypes, for example the topos of 'pro bono publico' ('to the advantage of all'), the topos of 'pro bono nobis' to the advantage of us'), and the topos of 'pro bono eorum' ('to the advantage of them').(Wodak, 2001, p. 74) In the media corpus, civic orientation is justified by media commentators as a useful initiative on the part of the Swedish state to the advantage of migrants (pro bono eorum) and to the advantage of all (pro bono publico) because it educates migrants on a variety of civic principles and values (including gender equality), which they are assumed not to already know.The assumption here is the learning of such principles and values will make them (migrants) more like us (Swedes).
Crucially, 'men' (in general) and migrant men (in particular) are seldom singled out as the main cause of lack of gender equality but appear in the neutral word cluster 'gender equality between men and women.'Moreover, the word 'men' (män) is not among the most recurrent words in the media corpus (N: 132), over twice less frequent than 'women' (kvinnor) (N: 307).From an intersectional perspective, however, it is interesting to note that the identity bundle of ethnicity and gender (migrant men) is the absent presence in those contexts where gender equality is overtly thematised in relation to violent patriarchal practices of arranged unions and child marriages.
The intersectional nexus point of ethnic masculinity is more overtly visible in the investigator's proposal for how civic orientation should be organised (SOU, 2010).The word 'man/men' (man/män), which is used to indicate migrant men in particular in this policy document, occurs six times: once, together with 'women' (kvinnor), in a general statement about the heterogeneity of the migrant population in Sweden; twice in statistical overviews of the gendered distribution of migrants from Thailand; and three times in sentences that display interesting transitivity patterns.In these cases, men are presented, together with women and children, as the victims of crime within their families ('it is the children, women, and men who are exposed to crimes within the families in their new home country') (SOU, 2010, p. 26); they are the implied agents in so-called 'bride import', a problem that 'led the government to consider introducing new legislation and clear guidelines about what women have the right to know regarding the background of the men they marry' (SOU, 2010, p. 26); or they are a hindrance to women's participation, and it is therefore proposed that there should be women-only civic orientation groups, which would allow for better inclusion of women 'whose relatives oppose their participation in groups where men are present' (SOU, 2010, p. 34).
That 'women' (kvinnor) are portrayed as passive and vulnerable can be found in the report about civic orientation (SOU, 2010) where they are represented as outside the public sphere, being at home because they have children directly after arriving in Sweden, and therefore 'at risk of falling outside the institutions and systems where they could acquire knowledge of the society in which they live' (SOU, 2010, p. 25).
Here, the underlying assumption is that migrant women have not been socialised into an egalitarian society as the Swedish one, from which they should learn.While it is unclear in these instances who is to blame, an analysis of the word husband(s) illustrates how migrant men are overtly singled out as the main agents causing migrant women's isolation and vulnerability.Firstly, it is argued that migrant women are not easily reachable by relevant information because 'their husbands do not allow them to participate in public events' (SOU, 2010, p. 31 emphasis added).Secondly, in reference to a report by the National Organisation for Women's Shelters and Young Women's Shelters in Sweden, it is emphasised that migrant women should be informed of their rights because 'harassment and violence are crimes in Sweden and their husbands cannot easily get them deported from Sweden' (SOU, 2010, p. 26).Admittedly, the words for 'husband(s)' do not specify the ethnicity of these men.However, the mutual constitution of ethnicity and masculinity can be inferred from the references to deportation and control of migrant women's access to the public sphere.Female migrant vulnerability is also a key ingredient in the media corpus, where 'women' (kvinnor) is one of the 40 most frequent content words (N: 307) (see Table 2).As noted above, ethnicity is seldom overtly thematised in the policy documents about civic orientation; it is the implicit, mutually constitutive signifier of gender to the point that, when men or women are mentioned in these documents, they refer specifically to non-Swedish men or women.In contrast, adjectives indicating nationality (foreign and foreign-born), and by proxy ethnicity, have been employed in media texts as overt qualifiers of 'women' in the corpus, always in opposition to 'native-born' (inrikesfödda) in statistical comparisons in which migrant women are over-represented in relation to unemployment.
Moreover, what can be seen here is a transitivity pattern in which 'women', together with children and young people (of both genders), are the patients of a variety of processes.A closer look at the contexts of usage of the passive forms utsätts (are exposed) and drabbas (are affected) indicates that the agents of these processes are not overtly named.However, the references to violence, harassment, honour killings, and forced marriages enable readers to infer who the agents of these patriarchal practices are (see Lingle (2022) about 'agentless passives' and the ability to infer agents from the co-text of a newspaper article).
Overall, the collocates of 'women' seem to indicate the presence of a topos of numbers about migrant women's poor integration in the Swedish labour market and a concomitant topos of danger about their vulnerability.This is used to justify the intervention of the state in the form of compulsory civic orientation courses, which are presented as the panacea of gender imbalance in Sweden.The characteristics of such a remedy come into relief if we consider the collocates of the two Swedish words for 'value(s)': värde(n) och värdering(ar).According to the Swedish dictionary (Svenska Ordboken), värde(n) indicates a more objective measurement of demand, utility or size, whereas värdering(ar) refers to more subjective perceptions.
Interestingly, värderingar is never used in policy documents about civic orientation.However, värde(n) occurs once in the Government's directive (Dir., 2009) and seven times in the investigator's report (SOU, 2010).In all these occurrences, värden collocates very strongly with 'fundamental' (grundläggande), as in the expression 'fundamental values', which refers broadly to 'democracy and equal value of all people' (SOU, 2010, p. 14) (see also Excerpts 1 above).Here is it important to point out that the investigator's report explicitly cautions that 'it can be difficult to specify what is meant by a society's fundamental values' (SOU, 2010, p. 17).In the light of this definitional quandary, the report clarifies that 'our stance on the matter is to take the [Swedish] constitution as a starting point,' and goes on to stipulate that the content of civic orientation should encompass 'knowledge of practical nature as well as [information about] institutional and organisational aspects of democracy and the welfare state' (SOU, 2010, p. 17).
A different distribution characterises the media corpus, where värdering(ar) is among the 40 most recurrent content words (N: 304), compared to värde(n) (N: 71).Like in the policy documents, värde(n) collocates strongly and frequently with words such as 'fundamental' and 'democracy'.In contrast, a more complex and detailed picture emerges if we consider the collocates of värdering(ar).
Table 3 indicates which values different media commentators believe migrants should learn in civic orientation classes.Here, the collocates of värderingar in the media texts partly overlaps with those of värden in the policy documents, giving the impression that these values are presented as universal, being rooted in the secular and liberal tradition of the Enlightenment, such as the upholding of a democratic state and the respect for human rights.Unlike in the policy documents, though, values shared across different polities in the world are particularised: they are explicitly tied to Sweden and Swedishness and are presented as Swedish national traits.Such an indexical link is not simply created by the strong collocation between 'values' and 'Swedish' / 'Sweden', but is also crafted in a more subtle way through the personal pronouns 'we' and 'our'.In this respect, as Billig (1995) noted, personal pronouns are 'small words' that play a key role in the discursive construction of an imagined national community.It is often difficult to establish precisely to whom 'we' and 'our' refer and thereby determine who is included or excluded from the 'imagined community'.Rather, 'we' and what counts as 'our' often (Petersoo, 2007) simultaneously include and exclude different referents.This is also the case in the examples at hand, where 'we' and 'our' may refer to the members of a particular party, everyone who lives in Sweden, ethnic Swedes only, or a combination of some or all these referents.Whatever the referent may be, there is an underlying understanding that gender equality is distinctively linked to Sweden; as the then-leader of the centre-right Moderate Party, Anna Klingberg Batra, explicitly stated in the columns of the conservative daily Svenska Dagbladet: 'integration is not just about work.Understanding and respect of Swedish values such as gender equality and individual's freedom are equally important' (2016-07-09).This statement is not very dissimilar to what a politician of a diametrically opposite political party, the Social Democratic Stefan Löfven, had said a few weeks before (see Introduction above).Another interesting point in the collocates of values is a tension between a discourse that requires a one-way process through which migrants adapt to a pre-given set of Swedish values, and a reciprocal dialogue about Swedish society.As the section below shows, it is this tension that lies at the heart of civic orientation classes.
In sum, collocate analysis of a corpus of mainstream media data and policy documents about civic orientation has enabled us to map the pervasiveness of discourses that construct two main intersectional identity bundles in which ethnicity and gender mutually constitute each other; namely, that of migrant women as a vulnerable group who lack access to the public sphere and the labour market.Migrant men are seldom overtly singled out as the cause underlying women's vulnerability; rather, they are the absentpresence that can be inferred from the context.Against this backdrop, values such as gender equality, respect for young people's rights, and freedom are presented as things that migrants do not already know but need to learn through civic orientation classes.What is most remarkable for the purpose of this paper is that such values are nationalised; they are presented as something typically 'Swedish'.Are we stating the obvious?Working on discourses of unemployment, Mautner (2009, p. 128) argued that Detractors of corpus-based methods could argue, of course, that one hardly needs a huge database of text and sophisticated software to 'prove' that being unemployed is not a pleasant thing.On the other hand, we should not forget […] that a fair proportion of any empirical work is devoted, precisely, to finding evidence for the intuitively obvious.
Therefore, while our findings may appear purely descriptive and obvious, they provide a robust empirical grounding for the results of previous qualitative investigations about Swedish political discourse (see, e.g. de los Reyes, 2016; Carlson et al., 2021;Mulinari, 2021) that demonstrated the existence of intersectional discourses that mutually constitute gender and ethnicity, representing (1) the 'migrant man' as threatening, (2) the 'migrant woman' as passive and in need of rescue, and (3) gender equality as an inherently Swedish 'national (essential) attribute' (de los Reyes, 2016, p. 28), to which migrants are expected to adapt.What remains to be understood is whether, and if so how, such discourses are reproduced and/or contested in actual interactions in civic orientation classes.

Gender equality and ethnicity in (inter)action in civic orientation classes
All the civic orientation courses that we followed reiterated that gender equality is a significant topic that migrants need to learn.On one occasion, during a recap of topics of the previous session, the communicator emphasised: 'And the most important topic!Jämställdhet [gender equality]' (Fieldnotes, May 2020).In another course, the participants were asked to vote on what they considered the most important component in a democratic society from among the following alternatives: (a) religious freedom, (b) gender equality, and (c) children being given space to participate in family issues.Four participants voted for (a), eight for (b), and only two for (c).On another occasion, the communicator shared a personal story about how he earned a higher salary than his female colleague in his country of origin, even though she was more qualified than he was.He then asked the participants if such gender inequities happen in Sweden.'Yes', answered a participant, which the communicator firmly rebutted saying: 'It is forbidden here' (Fieldnotes, April 2020).These examples not only illustrate the central place occupied by gender equality in civic orientation classes but also provide an indication of the mutual constitution of gender and ethnicity/nationality: Sweden as a gender equal paradise vis-à-vis gender-unequal contexts from what is problematically presented as a homogenous 'Arab world'.We consider this dual intersectional construction of Sweden vs. the 'Arab world' in more detail now in relation to family responsibilities, domestic violence and honour crimes.
In civic orientation courses in Arabic and English, it was not uncommon for participants to be given advice on how to achieve a successful marriage.As one communicator put it, when addressing male participants, '[i]f you want to have a good relationship with your wife you have to do at least 50% of the tasks at home' (Fieldnotes, April 2020).They were also told to 'prioritise your relationship, do fun things together' (Fieldnotes, April 2020) and talk to each other about problems; jealousy, by contrast, is not romantic.All these recommendations are built on several problematic assumptions: (1) a heteronormative assumption according to which the participants in civic orientation classes are or aspire to be in heterosexual marriages; (2) a cultural assumption that migrants are jealous, do not openly discuss family issues, and do not share family chores equally, unlike Swedes, who are instead assumed to be the benchmark of a successful gender equal union.Why migrant spouses do not help each other at home was discussed in one of the Arabic language courses, where a female participant from Syria, Basma, said: I think it is a matter of inherited traditions!If you see that your father is not helping your mother at home, you will do the same.Even if you came to Europe.In Syria, a large number of women used to stay home and take care of children while men worked outside.When we came here, we could see that this is not the case, and it is difficult for some men to accept this new situation.(Fieldnotes, April 2020) From an intersectional perspective, it is interesting here how problems related to gender (in)equality were explained as the result of ethnically specific socialisation processes, which became exacerbated after moving to Sweden, a context in which entrenched expectations about specific gender norms began to be questioned.Later in the same discussion, the communicator argued that this schism could be due to the fact that, 'In Arab countries, no one will say to women: you have to work or study or you have to be active.In Sweden, the government is behind every woman, and it is responsible to send her to work and study' (Fieldnotes, April 2020).Against this backdrop, participants were encouraged to discuss these cultural differences within their relationship; they were told it was their responsibility to do so.When all these examples are considered together, it is possible to see how an opposition is discursively created between Sweden (and an assumed undifferentiated category of Swedes), who are gender-equal, and the 'Arab world' and Arabic-speaking migrants, who are not, and need to learn about it.The achievement of such a gender-equal order is presented as combination of (1) state feminism according to which the Swedish state acts as a warrant of gender mainstreaming and supports women's participation in the public sphere, and (2) a more neoliberal logic of individual accountability for the achievement of future goals.The discursive opposition between gender in Sweden and gender in the 'Arab world' became even more tangible in discussions about honour killings, female genital mutilation and domestic violence.However, as we will see below, these were also topics through which new oppositions were refracted and identity boundaries were contested.
Various communicators broached the topic of honour crimes by talking about the organisation 'Never forget Pela and Fadime', an NGO whose name is a tribute to Pela Atroshi and Fadime Sahindal, two women of Kurdish descent who were killed in 1999 and 2002, respectively.In an English-language course, the communicator began by describing the murder of Pela Atroshi as follows: Excerpt 4 She was shot down in her family's house during a vacation trip to her own country.The motive was that she didn't live under the right rules.It was the first honour-related murders, the uncles where accused.They blamed the father.And later the uncles were sentenced to prison.(Fieldnotes, May 2020) It should be noted here how the murder was presented spatio-temporally as the first honour crime in Sweden, implying that such crimes had never occurred before in the Swedish context.While we do not wish to downplay the reality of honour-based crimes, their construction as purely a 'foreign' phenomenon that had never happened before in Sweden and was imported with increased migration is at variance with the otherwise detailed description of the meaning of honour given by the communicator.In a historical analysis, Lindstedt Cronberg (1997) showed how heder och ära (lit.honour and glory) were key components in the patriarchal structure of Swedish peasant society, disciplining women's sexuality between the 17th and 19th centuries.While there are no documented examples of a family murdering a woman who defied the honour and glory of the time, these patriarchal sexual values did not characterise a specific group, but permeated the entire state apparatus, which legislated, for example, that 'death was a possible punishment for so-called double fornication (when both parties were married) and in some cases for single fornication (when one of the parties was married)' (Österberg, 2005, p. 179).According to Österberg (2005), the Swedish state, through legislation, maintained control on women based on honour and shame until the 1930s.In such a system, 'the norm was defended with draconian punishments by the state and the state church, what was the need for the family to further discipline those who had already been punished?'(Österberg, 2005, p. 181-182).
In the light of these observations, it could be argued that, whether enforced by the family or the state, honour is a cross-cultural phenomenon that stems from a patriarchal belief that men have the right to control women's bodies and behaviours.Interestingly, however, there is a tendency in Swedish public discourse to portray domestic genderbased violence committed by Muslim and other migrant men against women as 'honour crimes' (hedersbrott) (Ekström, 2009) while similar offenses perpetrated in the home by Swedish men are minimised or even whitewashed in the media through euphemisms such as 'family drama' (familjdrama) or 'family tragedy' ( familjtragedi) (Dahlgren, 2014).Through such apparently mundane practices of labeling, honour is tied to specific ethnicities, cultures and/or religions while the behaviour of Swedish men is represented as less patriarchal and ultimately less objectionable.
The participants in the English courses were familiar with the notion of honour crimes and shared their experience of gender-based violence in their countries too, being 'common in the Balkans, blood-revenge.It's old traditions and within different tribes' (Fieldnotes, April 2020).However, the communicator was quick to bring back the discussion to 'Arabic cultures'.Of course, this could be simply an idiosyncratic choice of the communicator, who was from the Middle East, but it is striking that in an English-speaking course the examples of honour crimes were taken from the 'Arab world', and there was no detailed discussion of similar phenomena in the participants' contexts.
The discussions in the Arabic-language courses were more detailed: Excerpt 5 M [the communicator] explained examples of honour crimes like when you force your daughter to travel somewhere and get married to someone.M added that Sweden is suffering from this kind of violence and from violence between relatives.M added and God be with him Ya wilū (warning of horrible consequences) that who proves to be participating in an honour crime whether directly or by hiding information about it (Fieldnotes, April 2020) Here, detailed examples were provided, and the participants were expected to know and relate to them.In another course, a communicator said: 'I am a woman, and even if I am not personally subjected [to honour related oppression] there are those around me who are.It is important that we talk about and stand up to it' (Fieldnotes, April 2020).The participants were then presented with a specific example: 'Keeping someone locked up and isolating them for seven months as a punishment is forbidden, but it does not matter if it is done by a relative' (Fieldnotes, April 2020).While it is certainly not legal to keep someone imprisoned for months, it is worth noting that these discussions only occur in the Arabic-language civic orientation courses.In the English-language courses, the communicator explained at length what honour means (Fieldnotes, April 2020), something that was taken for granted in the Arabiclanguage courses.Overall, these examples reveal the centrality of the family and its importance as an organising element of socialisation in civic orientation courses.This is because the family is assumed to be important for migrants and it is the locus where the distance between 'their' values and Swedish culture is assumed to be more striking.Moreover, the greater volume of discussion in the Arabic-language classes on how to organise the family and deal with problems associated with domestic life indicates a controversial construction of the 'Arab world' as antithetical to Sweden, while English speakers are viewed as less problematic.However, such clear-cut dichotomies were sometimes questioned, such as in the cases when domestic violence was discussed.
In one of the Arabic-speaking groups, the communicator presented domestic violence as a major problem, but qualified the statement by adding that 'this violence is not only apparent in Arab countries, but even in Sweden when men are probably drunk' (Fieldnotes, April 2020).This represents an attempt to delink gender violence from Arab contexts.However, in the case of Sweden, it is explained away as the result of men's drinking habits rather than as a deliberate act or an effect of power structures.Note the usage of 'probably', which hedges the speaker's epistemic commitment to the propositional truth of the statement.That gender violence can be committed by Swedes was also raised in an exchange in one of the English courses, where the participants had to convince the communicator that women can also commit violence against men.A participant told the story of an acquaintance who had moved to Sweden from an African country and had been abused by his Swedish wife, but the authorities did not believe him (Fieldnotes, April 2020).The tension between Sweden and the 'Arab world' also transpired on several occasions in an Arabic-language group in a discussion about control and forced marriage.The communicator told the participants that some girls are not allowed to move or go out before they are married, and that the family enforces this control.A male participant from Kurdistan, Kawa, raised his hand and said 'but Sweden is better, there are laws here which protect and give more freedoms and rights' (Fieldnotes, March 2020).Here, the communicator reminded him that it is not a matter of better or worse, but about different perspectives, implying that there are problems in Sweden as well.On another occasion: Excerpt 6 The communicator asks Fouad what he thinks about the fact that Sweden is not completely equal when it comes to the labour market and health.Fouad says that he has not heard anything about that, that wages are not equal.The communicator replies that 'as [researcher] said, it is not 100 percent equal in Sweden'.Fouad responds that well, but in comparison to our countries it is so much better here, so I don't really understand your question.(Fieldnotes, April 2020) Not all participants agreed.A man from Syria, Ra'ed, protested that the level of equality differs greatly between urban and rural areas in his country, and he did not agree that everything is better in Sweden.He further claimed that it depends a lot on personalities and the individuals, and that it is important to have mutual understanding and to share the workload.Overall, these examples testify to momentary punctures in the discursive construction of opposing monolithic spheres -Sweden versus the 'Arab world'and Ra'ed's intervention is illustrative of a new boundary between urban and rural.
Such a problematisation of differences within the Arabic-speaking contexts also arose in connection with a discussion about female genital mutilation.Ra'ed asked if such a practice also occurs in Sweden: Excerpt 7 No, the communicator replies, there is no female genital mutilation, but it has also been illegal since a few years back.In Sweden it is foreigners who practice it.Over the summer they take their daughters to African countries, for example Sudan.(Fieldnotes, May 2020) Analogous to Excerpt 4 above, we see here a spatio-temporal positioning that portrays female genital mutilation as something that does not happen in Sweden but is performed elsewhere (Africa) during the summer months.While the previous excerpts presented the 'Arab world' as a monolithic violent Other in relation to honour crimes vis-à-vis a gender equal and not violent Sweden, here the usage of the word 'foreigners' [ajānib], together with the personal pronoun 'they' is indicative of a discursive boundary created by the communicator within the Arabic-speaking world and a discursive act of distancing of 'us' (including the communicator and the participants in the class) from 'them'; in this case, migrants of Sudanese origin.

Concluding remarks
This article offers an empirical contribution to ongoing debates about the incorporation of gender equality into civic orientation programmes for migrants.Taking Sweden as a case in point, our analysis of policy and media representations of civic orientation has demonstrated the existence of a dominant femonationalist discourse that presents gender equality as an inherently Swedish value, which needs to be taught to migrants, especially with a view to rescuing a supposedly homogenous group of migrant women from their dangerous male partners.However, ethnographic observations illustrate that such an intersectional bundle of gender equality and Swedishness is reproduced and contested in civic orientation classes.While Sweden is the presupposed 'good standard' of gender equality, as opposed to problematic representation of a homogenous, monolithic gender-unequal 'Arab world', there were also moments in which participants in civic orientation classes questioned such a dichotomy, provided details about the diversity CRITICAL DISCOURSE STUDIES between Arabic-speaking countries or within one and the same country, and discursively produced new boundaries, such as those between English-speaking migrants and Arabicspeaking migrants or within the Arabic-speaking community.This complexity notwithstanding, a bifocal lens on policy and media representations and ethnographic observations illustrates that when gender equality is invoked in the name of the state, the boundary between state feminism and femonationalism becomes razor thin.This raises a question about the role of critique within feminist scholarship.As feminist scholars, we strongly believe that feminism should inform state policies and practices.However, viewed from an intersectional perspective, discourses and practices of civic orientation in Sweden remind us that the co-optation of feminism into the state can be a tricky business that brings with it the risk of warping feminism into a trait of national/ethnic distinctiveness.Ultimately, our argument is that femonationalism is not the prerogative of far-right parties, but has already crept into the fabric of mainstream media and some educational practices aimed at migrants in a feminist state like Sweden.Van Dijk, T. (2000).On the analysis of parliamentary debates on immigration.In M. Reisigl, & R.

Table 2 .
Top collocates of women.